The LTAR Cropland Common Experiment at the Texas Gulf.

Texas Gulf is one of the 18 regional sites that is part of the USDA-ARS Long-Term Agroecosystem Research (LTAR) Network and focuses on cropland and integrated grazing land research in Central Texas, addressing challenges posed by soil characteristics, climate variability, and urbanization. This paper provides brief site descriptions of the two Cropland Common Experiments being conducted in the Texas Gulf LTAR region, emphasizing conservation tillage practices and precision agriculture techniques. The plot-scale study is located in Temple, TX, at the USDA-ARS Grassland, Soil and Water Research Laboratory and examines conventional tillage, strip tillage, and no tillage practices. The field-scale study, located in Riesel, TX, at the USDA-ARS Riesel Watersheds, assesses the impact of no tillage, cover crops, fertility management, adaptive management, and precision conservation on crop yield, profitability, and environmental footprint. Key measurements include soil and plant analyses, greenhouse gas fluxes, runoff water quantity and quality, and field operations recorded with precision agriculture equipment. Despite challenges posed by urban encroachment, future research aims to incorporate new technologies, such as unmanned ground vehicles, to enhance sustainability and productivity of the agricultural landscape. These experiments provide valuable insights for stakeholders, contributing to the development of sustainable agricultural practices tailored to the unique challenges within the Texas Gulf LTAR region.

Cropping management in a livestock-pasture-crop integration modifies microbial communities, activity, and soil health score.

Understanding indicators of soil health is crucial for developing agricultural systems that are sustainable and climate resilient. Labile soil carbon (C), microbial properties, and nutrient status are all incorporated into the Haney Soil Health Tool with the goal of summarizing several indicators into one index. Monthly soil samples from an integrated crop-livestock system in Central Texas were collected from 2017 to 2019. Fields represented a range of management practices, including cover crops, no-till, rotational grazing, and a native prairie remnant. Soil samples were analyzed for total C, water-soluble C, macro- and micronutrient content and bioavailability, and phospholipid fatty acids (PLFAs). Microbial activity was determined via a 24-h CO2 incubation. Soil health score, C, and PLFAs were well correlated with each other. The greatest total PLFA (219.5 nmol g-1 soil) and organic C (54.3 g kg-1 soil) were found in the native prairie, and the lowest were found in the unfertilized continuous-corn system (60.5 nmol PLFAs g-1 soil and 24.0 g organic C kg-1 soil). Of all agroecosystems, the perennial grazing system (soil health score, 24.7) was most similar to the native prairie (soil health score, 27.4), having high soil C and a large microbial community. Of the row cropping systems, the no-till system approached the perennial systems better than the conventional till and unfertilized conventional till (soil health score, 11.1 vs. 8.0 and 5.3, respectively). This study highlights the value of perennial grass grazing in agroecosystems and appropriate best management practices. Expanding this analysis to other sites may provide additional insight.

Predicting carbon and water vapor fluxes using machine learning and novel feature ranking algorithms.

Gap-filling eddy covariance flux data using quantitative approaches has increased over the past decade. Numerous methods have been proposed previously, including look-up table approaches, parametric methods, process-based models, and machine learning. Particularly, the REddyProc package from the Max Planck Institute for Biogeochemistry and ONEFlux package from AmeriFlux have been widely used in many studies. However, there is no consensus regarding the optimal model and feature selection method that could be used for predicting different flux targets (Net Ecosystem Exchange, NEE; or Evapotranspiration -ET), due to the limited systematic comparative research based on the identical site-data. Here, we compared NEE and ET gap-filling/prediction performance of the least-square-based linear model, artificial neural network, random forest (RF), and support vector machine (SVM) using data obtained from four major row-crop and forage agroecosystems located in the subtropical or the climate-transition zones in the US. Additionally, we tested the impacts of different training-testing data partitioning settings, including a 10-fold time-series sequential (10FTS), a 10-fold cross validation (CV) routine with single data point (10FCV), daily (10FCVD), weekly (10FCVW) and monthly (10FCVM) gap length, and a 7/14-day flanking window (FW) approach; and implemented a novel Sliced Inverse Regression-based Recursive Feature Elimination algorithm (SIRRFE). We benchmarked the model performance against REddyProc and ONEFlux-produced results. Our results indicated that accurate NEE and ET prediction models could be systematically constructed using SVM/RF and only a few top informative features. The gap-filling performance of ONEFlux is generally satisfactory (R2 = 0.39-0.71), but results from REddyProc could be very limited or even unreliable in many cases (R2 = 0.01-0.67). Overall, SIRRFE-refined SVM models yielded excellent results for predicting NEE (R2 = 0.46-0.92) and ET (R2 = 0.74-0.91). Finally, the performance of various models was greatly affected by the types of ecosystem, predicting targets, and training algorithms; but was insensitive towards training-testing partitioning. Our research provided more insights into constructing novel gap-filling models and understanding the underlying drivers affecting boundary layer carbon/water fluxes on an ecosystem level.